homericintelligence/scylla/analyze-equations

Analyze Equations

Extract and parse mathematical equations from research papers to understand computational requirements and implementation details.

When to Use

• Converting paper formulas to code implementations
• Understanding algorithm mathematical foundation
• Identifying performance-critical computations
• Planning SIMD optimization strategies

Quick Reference

# Extract LaTeX equations from PDF
pdftotext -layout paper.pdf - | grep -E '\$\$|\\begin\{equation\}' | head -20

Workflow

1. Extract equations: Identify and extract LaTeX/mathematical notation from source documents
2. Parse components: Break down complex equations into primitive operations
3. Identify variables: Document input parameters, intermediate values, output
4. Determine data types: Specify scalar vs vector operations, precision requirements
5. Map to implementation: Connect mathematical notation to code operations (matrix multiply, activation, etc.)

Output Format

Mathematical analysis document:

• Extracted equations (with source references)
• Component breakdown (operands, operations)
• Variable definitions and constraints
• Data type requirements (float32, float64, int32)
• Mojo implementation mapping

References

• See CLAUDE.md > Language Preference for Mojo ML implementations
• See extract-algorithm skill for algorithmic interpretation
• See /notes/review/mojo-ml-patterns.md for Mojo tensor operations